Universal three-drug antiretroviral therapy (ART) for HIV-infected women offers many advantages in terms of preventing HIV transmission to children, safer breastfeeding, improved maternal health, reduced sexual transmission of HIV, and simplified treatment programs. Safety of increased ART exposure during pregnancy and conception for mothers and their children is not fully established, and there are concerns about maternal adherence and program effectiveness. Investigating the impact of these changes is complex due to weak clinical and information systems in many settings, rapidly changing guidelines and service realities, and difficulty ensuring long term follow-up of sufficient numbers of children to exclude potential harmful effects. The current HIV antenatal seroprevalence in the Western Cape province of South Africa is 19% with vertical transmission of HIV estimated at 2%. Providing maternity care to 100,000 mothers a year, the province has a long history of innovation in PMTCT programs; recently introducing life-long universal ART for HIV- infected pregnant or breast-feeding women. The province is uniquely placed to comprehensively evaluate this guideline change due to the availability of electronic data on the ART program, and the ability to link via a unique health identifier to a range of other datasets to improve follow-u, ascertainment of pregnancy and ART exposure, laboratory values, vital status data and other outcomes. This proposal seeks to evaluate the impact of the guideline change on: 1) the effectiveness of PMTCT and ART programs for participating women based; 2) the impact of increased ART exposure antenatally and around conception on maternal and perinatal health outcomes; 3) the impact on child health outcomes for children exposed in utero to ART; and finally 4) the impact on the governmental ART program and on the epidemic itself. At each level of evaluation, a population-based province-wide activity is paired with a sentinel site activity which probes clinical and operational issues in greater depth. All of the sentinel site activities are located in the same sites to improve operational efficiency, while all of the province-wide data activities will be undertaken by a single team. Program effectiveness will be assessed both through cohort analyses of the entire provincial ART cohort of 35,000 adult patients starting ART each year, and a detailed cohort study of 900 women in the sentinel site. Maternal and perinatal health outcomes will be assessed through provincial mortality reporting and perinatal audits which both have high levels of completeness, complimented by participation at the sentinel site in the national pregnancy exposure (PER) register and birth defects surveillance register. Post-neonatal child health outcomes will be tracked through hospital admissions, which for children from the PER, will be reviewed to verify discharge diagnoses. Program and epidemic impact will be explored through mathematical modeling based on a new country level HIV and demographic model.